The Hive Mind is a Single Reinforcement Learning Agent

Decision-making is an essential attribute of any intelligent agent or group. Natural systems are known to converge to effective strategies through at least two distinct mechanisms: collective decision-making via imitation of others, and trial-and-error by a single agent. This paper establishes an equivalence between these two paradigms. We show that the emergent distributed cognition (sometimes referred to as the hive mind) arising from individuals following simple, local imitation-based rules is that of a single online reinforcement learning (RL) agent aggregating action-value samples from many parallel instances of a multi-armed bandit. More specifically, we show that, in the blindly imitative weighted voter model of honey bees' waggle dance, the update rule through which this macro-agent learns is an RL algorithm that we coin Maynard-Cross Learning. Our analysis implies that a group of individuals following simple imitative strategies can be equivalent to a more complex learning entity, substantiating the idea that group-level intelligence may explain how seemingly irrational individual behaviors are selected in nature. Beyond biology, the framework offers new tools for analyzing economic and social systems where individuals imitate successful strategies, effectively participating in a collective learning process. Our findings may further inform the design of scalable RL-inspired collective systems in artificial domains.